Asmae Amarir

Key Skills

Software

Appen

Clickworker

Labelbox

OneForma

Top Subject Matter

No subject matter listed

Top Data Types

Document

Text

Video

Top Task Types

Classification

Data Collection

Emotion Recognition

Evaluation Rating

Translation Localization

Freelancer Overview

Multilingual linguist and language specialist with over six years of experience in translation, transcription, and AI training data evaluation. Skilled in linguistic quality assurance, prompt engineering, and audio-text alignment for AI models. Proven track record working on diverse language-related projects including machine translation post-editing (MTPE), multilingual transcription (French, Arabic, Moroccan Darija), and LLM subjective evaluation tasks such as audio dubbing accuracy, voice naturalness, and prosody similarity. Experienced in classification, data annotation, content moderation, and comparative evaluation of AI-generated outputs. Adept at applying structured guidelines and scoring systems (e.g., ACR, SBS) to evaluate audio and text data for AI training and fine-tuning. Passionate about improving AI systems' understanding of multilingual and culturally nuanced content.

Labeling Experience

Bilingual LLM Response Evaluator

LabelboxTextClassificationEvaluation Rating

As a Bilingual LLM Response Evaluator, I assessed and compared AI-generated responses in both English and French, judging their quality based on accuracy, fluency, and logical coherence. My role involved selecting the best response and providing detailed justifications for my choice, ensuring linguistic precision and cultural relevance in both languages.

As a Bilingual LLM Response Evaluator, I assessed and compared AI-generated responses in both English and French, judging their quality based on accuracy, fluency, and logical coherence. My role involved selecting the best response and providing detailed justifications for my choice, ensuring linguistic precision and cultural relevance in both languages.

2024

Translation Evaluation

AppenAudioAction RecognitionTranslation Localization

I evaluated AI-generated speech samples for emotion, intonation, rhythm, and voice similarity. Using the Halo tool, I compared audio pairs and rated them on a -3 to 3 scale for naturalness, translation accuracy, and prosody consistency. My role involved detecting subtle nuances in speech patterns, identifying issues like robotic tones or inconsistent emphasis, and rejecting samples with poor audio quality or irrelevant content. This work contributed to refining AI voice models by ensuring culturally and linguistically accurate synthetic speech.

I evaluated AI-generated speech samples for emotion, intonation, rhythm, and voice similarity. Using the Halo tool, I compared audio pairs and rated them on a -3 to 3 scale for naturalness, translation accuracy, and prosody consistency. My role involved detecting subtle nuances in speech patterns, identifying issues like robotic tones or inconsistent emphasis, and rejecting samples with poor audio quality or irrelevant content. This work contributed to refining AI voice models by ensuring culturally and linguistically accurate synthetic speech.

2024

Education

Work History